Best Evidence Topics
  • Send this BET as an Email
  • Make a Comment on this BET

Is haloperidol or a benzodiazepine the safest treatment for acute psychosis in the critically ill patient?

Three Part Question

In [patients in a high dependency setting] is [haloperidol or benzodiazepines] the safest treatment in [treating acute psychosis]?

Clinical Scenario

Its 1am and you are the registrar on call. You are asked to see a 70 year old lady 3 days post aortic valve replacement. She is confused and agitated and is trying to pull out her central line. She has already removed her arterial line and is on the side of the bed demanding to see her husband and wanting to get her clothes to go home. Her pulse is 160 and she is in AF, her BP is 100/55 and her oxygen saturation is 92%. She has no past psychiatric history and the nurses are very keen to see her sedated as it is currently taking 3 people to keep her under control. You want to control her agitation but you are anxious that you may oversedate her or cause cardiac side-effects as a result of additional drug administration. You wonder what is the optimal policy for sedation of this agitated lady.

Search Strategy

Medline 1966-Nov 2003 using the OVID interface
[exp haloperidol OR OR OR exp lorazepam OR or exp diazepam OR ] AND [exp Critical Care OR exp Critical Illness/ OR critically ill OR exp Intensive Care Units/ OR coronary care OR cardiac surg$ OR High]

Search Outcome

294 papers were found of which 5 papers were clinically relevant. One of these papers summarised several case reports that were also found. In addition 4 more papers were found on checking reference lists. All these papers are presented in the table

Relevant Paper(s)

Author, date and country Patient group Study type (level of evidence) Outcomes Key results Study Weaknesses
Hassaballa HA & Balk RA,
Review of papers documenting torsades de pointes after iv haloperidol 19 individual cases of torsades de pointes found in the literatureReview of case series (5)Dose required to cause torsades de pointesMean dose 507mg. Lowest dose single dose of 10mg. Highest dose 1,000mg in 30 minsSelected collection of individual cases from the literature Search strategy for finding papers not given
Outcome of torsades de pointed in these patientsNone of these patients died as a result of the episode
Tisdale JE et al,
30 critically ill patients who received intravenous haloperidol for delusional agitation 6 patients known to get torsades de pointes after haloperidol therapy 24 control patientsCase-control study (3b)QT interval in study ptsPrior to haloperidol 501 +/- 44 ms. After haloperidol 606 +/- 61 ms vs. p = 0.007Note a 3 page erratum section appears in a subsequent issue of J Clin Pharmacol, with several significant changes, compared to the original article
QT interval in control ptsPrior to haloperidol 466 +/- 44ms. After haloperidol 507 +/- 60 ms vs. p = 0.01
Odds of developing torsades12 fold increased risk in those with QT interval > 521 msec
Breitbart W et al,
Nondelirious, medically hospitalized AIDS patients were prospectively entered into the study. Randomization to treatment was performed if patient subsequently met DSM-IIIR criteria for delirium Treatment: haloperidol (N = 11), chlorpromazine (N = 13), lorazepam (N = 6).Double Blind PRCT (2b)Improvement in delirium score at 24 hours (mean initial score was 20)Haloperidol: 8 point improvement, chlorpromazine:8 point improvement, lorazepam: 1.5 point improvement. p<0.001 neuroleptics vs lorazepamNot post operative or cardiac surgical patients 5 patients died within 8 days of initiation of treatment Low doses of all drugs used, haloperidol mean 2.8mg, chlorpromazime 50.0mg, Lorazepam 3.0mg in 1st 24 hours Small numbers
Adverse effectsExtra-pyramidal side effects were low with haloperidol and chlorpromazine, but ALL patients with lorazepam had treatment limiting side effects including oversedation, disinhibition, ataxia, and increased confusion, causing study to be stopped early
Shapiro AJ et al,
Development of practise parameters on behalf of the American College of Critical Care Medicine Task force of 40 experts from the society convened to construct these guidelines in conjunction with full literature review over the course of 1 yearSystematic Review (3a)Haloperidol is the preferred treatment for delirium in the critically ill patientPresented as level 1 evidence. Usual dose 5-10mg. Onset of action 30-60 min after IV administrationNot specifically constructed for patients post cardiac surgery Quoted papers in support of recommendations are case series? and reviews and not well conducted clinical trials, calling level of recommendation into question
Benzodiazepines not recommended for acute deliriumStated that benzodiazepines may cause a paradoxical worsening of symptoms. No mention made of cardiovascular effects
Lorazepam is preferred agent for prolonged treatment of anxiety in the critically ill patientPresented as level 2 evidence. Lorazepam causes less hypotension than other benzodiazepines
Huyse F & van Schijndel RS,
Case report of a single patient on an intensive care unit for legionella pnumoniae pneumonia receiving haloperidol 7.5mg of iv haloperidol administered for agitation while weaningCase report (5)ComplicationPatient suffered a cardiac arrest immediately after haloperidol administration. A second cardiac arrest occurred 2 hours later immediately after a second 7.5mg dose of haloperidolSingle case report Not a patient post cardiac surgery
OutcomePatient discharged home
Tesar GE & Stern TA,
Review of the literature and presentation of a protocol for tranquilisation of Agitated ICU patients Protocol : 0.5-2.0mg mild agitation 5.0-10mg moderate agitation 10mg or more severe agitation Allow 20 mins before repeat dose If agitation persists double dose and repeatReview (5)Haloperidol use and complicationsOnset of action 10-30mins but is very safe. Haloperidol produces trivial effects on haemodynamic functionReview is essentially expert opinion rather than systematic review of the literature
Additional use of lorazepamLorazepam has fewer cardiovascular complications than other benzodiazepines. In severe agitation lorazepam 2-10mg may also be given
Adams F,
Retrospective cohort of 2,000 medically ill patients with cancer Protocol used : 5mg of haloperidol and 0.5mg of lorazepam and hydromorphone. 20 mins later if no response 10mg haloperidol and 0.5-2mg lorazepam. Repeat this at 30 min intervals until agitation controlledCohort study (4)Clinical experience in 2000 patientsProtocol has been safely used in 2 cancer centers for 8 yearsNo clinical data or follow up data given about these 2000 patients No demographics or patient selection given Safety seems to be decided according to authors clinical memory Not cardiac patients
Rao S et al,
36 patients attending for endoscopy received intravenous diazepam. Respiratory and cardiopulmonary function was measured using indicator-dilution technique with CVP and invasive BP monitoring 18 patients has moderate to severe obstructive lung disease received mean 34mg diazepam 18 patients with normal lung function received mean 58mg Enough diazepam was used to allow endoscopyCohort study (2b)Cardiac parameters post diazepam in normal patients (Pts with abnormal lung function had similar or less marked findings)Heart rate mean inc 13.4%. Blood pressure mean drop 12.5%. Stroke volume mean drop of 31.5% at 15 mins. Cardiac output mean dec. 14%Non post operative, non cardiac patients High doses of diazepam
Respiratory parameters after diazepampCO2 mean 8% increase. SaO2 normal pts no change, lung pts mean drop 92%-88%
TesarGE et al,
Case report of 4 cardiac patients in the Coronary Care Unit requiring high dose Haloperidol for control of agitation. Initial starting dose of 5mg but rapidly increased to single doses of 30-75mgCase Series (5)Dose of haloperidolMean dose of 100mg per day to control agitation. Patient 1 had 140mg on day 1 then 7.5mg per 4 hours maintenance. Patient 2 had 20mg per 4 hours and 270mg in one day. Patient 3 had 485mg over 8 hours. Patient 4 had up to 530mg per daySmall case series
ComplicationsNo complications reported in these 4 patients


The American College of Critical Care Medicine has provided the highest quality review in this area. They conclude that the first line drug for acute agitation in the critical care environment is haloperidol and that cardiovascular side effects are rare. In addition they state that benzodiazepines may in fact exacerbate symptoms, although lorazepam is a safer second line drug than other benzodiazepines such as diazepam. However this guideline is mainly based on expert consensus in conjunction with case reports. Rao et al demonstrate the unacceptable cardiovascular side-effects with diazepam with an average 14% drop in cardiac output. Breitbart et al in a well conducted randomized controlled trial compared lorazepam and haloperidol in medically unwell AIDS patients. They found that haloperidol produced significantly better resolution of delirium and lorazepam invariably produced treatment limiting side-effects including increased confusion , disinhibition and ataxia. We could find no other papers that documented a poorer performance of lorazepam compared to haloperidol either in terms of side effects or resolution of delirium, and in fact Adams et al described the safe use of a combined policy of both drugs given together. This utilises the fact that the onset of action for lorazepam is faster and allows a lower dose of haloperidol to be used. Tesar et al reported the safe use of haloperidol in doses of 100mg per day or more in agitated patients post cardiac surgery and presents a protocol for its safe use, but the complication of Torsades de pointes is well documented. Hassaballa reported 19 cases of Torsades after haloperidol and Huyse reported a case of cardiac arrest after 7.5mg of haloperidol. Tisdale performed an interesting study that demonstrated that haloperidol always causes a prolonged QT interval and advocates cardiac monitoring and extra caution with a QT interval over 521 msec.

Clinical Bottom Line

Haloperidol should be considered the first line drug for agitated patients post cardiac surgery, however lorazepam either alone or in conjunction with haloperidol is an acceptable alternative.


  1. Hassaballa HA, Balk RA. Torsade de pointes associated with the administration of intravenous haloperidol. Am J Ther 2003;10(1):58-60.
  2. Tisdale JE, Rasty S, Padhi ID, et al. The effect of intravenous haloperidol on QT interval dispersion in critically ill patients: comparison with QT interval prolongation for assessment of risk of Torsades de Pointes. J Clin Pharmacol 2001;41(12):1310-1318.
  3. Breitbart W, Marotta R, Platt MM,et al. A double-blind trial of haloperidol, chlorpromazine, and lorazepam in the treatment of delirium in hospitalized AIDS patients. Am J Psychiat 1996;153(2):231-237.
  4. Shapiro BA, Warren J, Egol AB et al. Practice parameters for intravenous analgesia and sedation for adult patients in the intensive care unit: an executive summary. Society of Critical Care Medicine.[comment]. Crit Care Med 1995;23(9):1596-1600.
  5. Huyse F, van Schijndel RS. Haloperidol and cardiac arrest. Lancet 1988;2(8610):568-569.
  6. Tesar GE, Stern TA. Rapid Tranquilization of the Agitated Intensive Care Unit Patient. J Intensive Care Med 1988;3(4):195-201.
  7. Adams F. Emergency intravenous sedation of the delirious, medically ill patient. J Clin Psychiat 1988;49:(12 SUPPL. DEC);p 22-27.
  8. Rao S, Sherbaniuk RW, Prasad K, Lee SJ, Sproule BJ. Cardiopulmonary effects of diazepam. Clin Pharmacol Ther 1973;14(2):182-189.
  9. Tesar GE, Murray GB, Cassem NH. Use of high-dose intravenous haloperidol in the treatment of agitated cardiac patients. J Clin Psychopharmacol 1985;5(6):344-347.